The BIDS community is proud to announce the release of Common Derivatives! Derivatives are the outputs of preprocessing pipelines, capturing data and meta-data sufficient for a researcher to understand and (critically) reuse those outputs in subsequent pipelines. This specification can be found at: https://bids-specification.readthedocs.io/en/stable/05-derivatives/01-introduction.html
Common principles for derivative datasets:
Derivative datasets describe the contents of the dataset, the methods by which it was generated, and the original dataset(s) that were used to generate it. This metadata is mandatory at the dataset level and optional at the file level.
Derivative files must be distinguishable from raw files. A generic description (desc-) entity allows this if no other entity applies.
The spatial interpretation of data, if applicable, can be specified with a space entity and SpatialReference metadata.
Releasing the common principles for derivative datasets will enable further progress into our modality specific derivative initiatives such as, electrophysiological (https://github.com/bids-standard/bep021), PET, and diffusion MRI (https://github.com/bids-standard/bids-bep016). Our immediate attention will shift to focus on specifying anatomical and functional derivatives (Look out for the pull requests soon!). BEP 028: (provenance - https://github.com/bids-standard/bids-specification/pull/487) is a closely related effort that we hope to see progress quickly as well!
We would like to thank our incredible community of contributors and members that helped push this extension over the finish line!
Chris and Franklin - BIDS Maintainers